Rotating garnets

Abstract
The origin of snowball and sigmoidal inclusion patterns in porphyroblasts is discussed. Snowball garnets are peculiar to shear zones whereas sigmoidal patterns occur in porphyroblasts both in shear zones and on the limbs of folds. There are currently two models for the development of snowball garnets and these have been discussed extensively in the literature. We show that although the typical two‐dimensional snowball pattern can be produced by either model, the three‐dimensional inclusion patterns are model‐specific thus providing a distinguishing criterion. We have applied this criterion to all the available data and find that the classical model, which is dependent on the rotation of garnet relative to a single foliation, is applicable in all cases. Syn‐kinematic porphyroblasts on the limbs of horizontal normal folds generally show little rotation relative to geographical coordinates. What rotation they do show generally has the same sense as that of the host limb, but is less in magnitude. This has been used as evidence that the porphyroblasts have remained irrotational while the rocks deformed around them; the implication being that they were unaffected by vorticity associated with folding. This has been explained by claiming that the porphyroblasts are restricted in distribution to small domains of coaxial deformation path. We show that for reasonable deformation models of horizontal normal folds, porphyroblasts affected by vorticity will rotate little with respect to geographical coordinates and our results predict the commonly observed natural patterns. We conclude therefore that lack of rotation relative to geographical coordinates cannot be used to demonstrate that porphyroblasts have grown only in coaxially deforming domains; much less restrictive and more reasonable interpretations are possible. Consequently, the lack of rotation relative to geographical coordinates is more significant for fold modelling than it is for the garnet controversy.